The eddy covariance method is extensively applied in the air-sea interface turbulent flux measurements. The quality of the flux measurements depends greatly on the exact use of the correction methods and the marine meteorological conditions. Therefore some post correction processes and quality controls to the data are necessary. Based on about 7, months measurements of 2 eddy covariance systems aloft the coastal marine meteorological platform at Bohe, Guangdong Province, the effects of some correction methods, which are not running on the data logger, are analyzed. The results show: the effect of despiking is apparent only when the fluxes are low. Mostly, the changes of the fluxes (including U*, latent heat, sensible heat and CO2 flux) are less than 1% by despiking. Due to the influence of the sea surface wave field, the planar fit method is apparently better than the double rotation scheme. The spectral response correction can enhance the fluxes by 5%~8%, and the enhancement during daytime is slightly less than that during nighttime. The ogive test shows that the averaging period of 30 min is probably not enough during the stability transition periods in daytime. Finally, the turbulence stationarity test and integral turbulence characteristics test show that about 50% data are of high quality and can be used for basic research. Comparison of the flux measurements on two levels also confirms that the data quality of the Bohe station.